Study of Social Dialects; Role Ofrace

$Study of Social Dialects; Role Ofrace$

. 4D . S.

OCIIIIIPPT BB SCIPIE I ED 032 535 AL 002 094

Ely -Waitron. Walter A. . Social DialeCts from a Linguistic PerspeCtive: Assumptions. Current Research.and Future Directions. Pub Date Oct 69 Note -Sip.; Paper presented at the Center forApplied Linguistics Conferenceon Approaches to Social Dialects. Washington. DEC.. October 31to November 1. 1969. EDRS Price ME -S025 HC -S2.95 Descriptors -Descriptive Linguistics. Dialect Studies.Linguistic Theory. Negro Dialects. Nonstandard Dialects. Research Needs. cial Dialects. Sociolinguistics.Surveys This paper begins witha discussion of the assumptions basic to the study of both language and social dialects:verbal systems are arbitrary. all languagesor dialects are adequate as communicativesystems. they are systematic and ordered and learned in the context of thecommunity. A survey of current work and findings in dialect studies follows. In the lastpart of the paper. the author discusses research needs in the areas of (1) field techniques(size of sample necessary fora reliable study of social dialects; role ofrace. sex. and social class of the fieldworker in affecting the speech ofan informant; elicitation procedures whichcan get at judgments of the grammaticality of nonstandardstructures apart from judgments about social acceptability; andimportance of statistical calculations in comparing the various quantitative measurements thatare made). (2) descriptive studies (intonation. Black English in the South. nonstandardwhite dialects. age-grading. acquisitional studies of nonstandard dialects. andmore descriptive date on the role of sex in language). and (3) theoretical issues (wayin which observed linguistic variationcan be accounted for in a linguistic model ofdescription. extent to which a descriptioncan encompass more than one idiolect. and desr:riptive differencesbetween several types of language situations). (DO)

ell

I 4 A l U.S. NPARTNENT 01 MALIN. EDUCATION t WINK s., OffKE Of EDUCATION Walter A. Wolfram

Center for Applied Linguistics DNS DOCUMENT HAS 11111 MOWED UAW AS RECEIVED FROM THE PeN October, 1969 MON 01 OKAINIATION 01116INATIN6 ITPOINTS Of VIEW OR OPINIONS STATED 10 NOT IMAM REMO! °KIM OFFICE Of EDUCATION Lin POSITION 01POUCV.

Pe% SOCIAL DIALECTS Mal A LINGUISTIC PERSPECTIVE: O ASSUMPTIONS, CURRENT RESEARCH, AND FUTURE DIRECTIONS *

L&J I

Linguistic Assumptions

The research assumptions of linguistics inrelation to the study

of social dialects are derived from the anthropologicaltradition of

culturalrelativismWhen anthropologists at the turn of the century

reacted to the evaluative measures of their predecessorsin describing

non-western cultures, they set the stage for a similar viewof language

differences. American anthropologists such as Boas, Kroeber,and

Herskovitz insisted on viewing cultures descriptivelyrather than by

some yardstick of evolutionary development. Such an approach pre-

cluded classifying a language as "underdeveloped","primitive" or

inherently inferior simply because it was used in aculture devoid of

the technological implements found in western civilization:the notion

of "primitive" languages was denounced as a productof ethnocentrism

by socially and technologically superordinate cultures.

Descriptive linguists; then, simply adopted the sameassumptions

about language that anthropologists had maintained fornon-linguistic

aspects of cultural behavior. Even as anthropologists rejected the

Procrustean mold of western civilization in describingother cultures,

This is not to say that there is unanimous agreement among anthropologists about the extent to which cultural relativityis a philosophical, descriptive ormethodological prerequisite for anthropological study. For an explication of some of the controversy concerning cultural relativity, see Schmidt(1955).

*Paper to be presented at the Center for AppliedLinguistics Conference on Approaches to Social Dialects, October 31-- November 1,1969, Washington, D.C. - 2 - linguistic descriptions rejected the moldof the classical languages in

describing non-Indo-European languages.The opinion that languages have many different ways of expressing "underlying"logical operations

became the cornerstone of assumptions about language differences. At

first, these assumptions were relevant mainly to languages compared

across clear-cut cultures; later they becamerelevant to the comparisons

of speech differences for different social levels(subcultures, if you will) within the framework of a larger culture. It is within the latter

framework that we shall discuss the basic assumptions conerning social

dialects in their relation to standard English.

What then, is the explicit nature of these assumptions? In order

to discuss these assumptions, we must begir with the primitive assump-

tions linguists accept in their definition of LANUAGE.

One of the basic premises about the nature of language is that

verbal systems are arbitrary, established only byconvention.2Although

one cannot deny a certain degree of consistency inthe relation of lane

guage to the outside world, relationships betweenobjects and linguistic

signs are arbitrary.All languages are equally capable of conceptu-

alization and expressing logical propositions, but the particular mode

(i.e. grammar) for conceptualizing may differ drastically between lan-

guage systems. The linguist, therefore, assumes that different surface

forms for expression have nothing to do with the underlying logic of a

Nide (1964:47) notes that the arbitrary character of linguistic symbols refers to: (1) the arbitrary relationship between the form of the symbol and the form of the referent (2) the relationships between classes of symbols and classes of referents and (3) the relationship between classes of symbols and classes of symbols. -3..

sentence. There is nothing inherent in a given language variety which

3 will interfere with the development of conceptualization. This is not

to say that differences between the handling of logical operations may

never correlate with different social classes; however, onthe basis of

this premise, it cannot be related to language differences, since all .

language varieties adequately provide for expression of syllogistic

reasoning.

To those familar with the current interest in nonstandard English,

particularly Black English, it should be apparent that this assumption

does not coincide with the conclusions of some of the current projects

in the area. To suggest that Black English imposes certain cognitive

limitations on the logical operations of the Black English speakerand

to reject it as "illogical" is not generally taken seriously bylin-

guists. Yet, the work of Bereiter and Ehglemann,(1965,1966) proposes

such a view. Ultimately, such notions seem to be derived for a pre-

scriptive norm for language usage, although philosophical dictumsabout

the logical nature of certain rules of a language add a ring of authority

to such pronouncements.

To illustrate, one of the most cited examples of the inherent logical

foundation of standard 13.41 lish is the use of negatives with indefinites.

If a person uses a sentence such as Johndidso*, it is

3. One should be careul to note the distinction between 'interference in conceptual developinentq and the Uhorfian hypothesis, which maintains that 1an3uaGe categories predetermine particular concep- tualizations of tae external world. In the former case a value judgeoent is place on the adequacy of conceptualization, while the latter, no value judgment is made. - 4 -

understood negatively, but ifa person should use the sentence, John

didn't do nothing, it (Inn fully- be meantas a positive statement since

two negatives logicallymake apositive.4In this view, if aperson

uses the construction in a sentencesuch as John didn't do nothAE

because he was so laas heis using English inan illogical way.

Therefore the sentence does notmean what the speaker thought it meant.

The speaker apparentlymeans that John did not work, but by saying

12hugaLLAININm: he affix= that Johnactually did something.

Interpretations of this sort ignorea quite regular rule in Black

English (as well as in languagessuch as Spanish and Italian) which

states that whenyou have a negative sentence with indefinites,you

may add a negative element toevery indefinite (e.g. lie ain't never

had no trouble aboutnone of us Pullin' out no knife or nothin0. In

the underlying structure thereis only one negative, which is simply

realized on every indefinite.

Essential to understanding theunderlying proposition of the above

sentence is the distinction between"deep" end "surface" structure in 5 language. Deep structure is basicallya system of propositions which

4. This sentence could, ofcourse, be interpreted positively in a context such as He didn't dojust nothing; he was always busy. Usually, however, there isa strong stress on nothir to indicate this intention.

5. Althow.A the notion of deep andsurface structure in modern linguistics derived from the insightsof transformztional-generative grammar, any generative model of language willbe characterized by the recognition of thisdichotomy. For example, this notion is implicit in stratificationalgrammar, although the series of steps (i.e. howmany levels) and the mode for relating levels (howone gets from one level to another)may differ significantly from transformational-generativegrammar. . 5 .

interrelate in such a way as to express the meaning of the sentence,

while surface structure is realization of these propositions in terms

of the particular grammatical devices (e.g. linear ordering, grammat-

ical categories) of the language. The knowledge of language involves

the ability to assign deep and surface structures to an infinterange

of sentences, to relate these structures appropriately, and to assign

a semantic interpretation and phonetic interpretation to the paired

deep and surface structure. The failure to understand this relation

is, no doubt, responsible for some of the misinterpretation of non-

standard varieties of languages. We see, in the case of Blackthlish

multiple negation, that the basis for arguing that it is not logical

is found in the mistaken identity of a surface structure for a deep

structure.

Proclamations about the inadequacy of Black English as a nonstandard

language variety on logical bases, from a linguistic perspective, are attributed to a naive disregard for one of the primitive premises about

the nature of language. Yet, Bereiter maintains that a difference between the negative patterns of Black English and standard English is an indication that the black ghetto child is "deprived of one of the most powerful logical tools our language provides." (1965:199)Bereiter claims that a black ghetto child "does not know the word not" since his subjects did not regularly give him the form in negating a sentence such as This is not a book.The assumptions of Bereiter, however, reveal two misconceptions. In the first place, he has confused the inability of the student to give him the word not in a specific elicitation task with the child's unfamiliarity with the lexical item.Labar (1969), -6-

observes that many of the formal elicitation procedures in the context

of a classroom can be quite intimidating to the student andthe best

defense may be no verbal response at all.Intensive research on

the structure of Black English in Washington,D.C. and Detroit clearly

indicates that not is an integral part of BlackEhglish. Secondly,

Bereiter is apparently unaware that theuse other negative patterns

may serve the same purpose as not.Thus, a sentence such as This

ain't no book may communicate thesame negative pattern as not although

the structure of the sentence isdifferent. What is essential is not

the occurrence of a particular lexicalitem, or a specific syntactical

pattern, but the realization ofa particular type of underlying struc-

ture involving negation.Whatever deficiencies in logical operations

may or may not exist among black ghetto children, thesehave nothing 6 to do with language.

A second assumption of the linguist is theadequacy of all languages

or dialects as communicative systems. It is accepted as a given, that

language is a human phenomenon whichcharacterizes every social group,

and that all language systemsare perfectly adequate as communicative

systems for the members of the socialgroup. The social acceptability

of a particular language variety (consideredthe nonstandard variety

because it is associated witha subordinate social group) is totally

unrelated to its adequacy for communication. The question concerning

different language varieties is not theWHAT but the HOW of communication.

Thus, the consideration of the so-calleddisadvantaged child as "non- verbal", "verbally destitute",or at best, "drastically deficient" in

It is to note thata sample of language indices Bereiter cites as indicative of languagecompetence (1965:199-200) have nothing to do with language. He consistentlycorfwes. therecognition of logical operations withlanguage development. - 7 -

his speech is diametically opposed to this basic assumption. That there

are typical situations in which young children do not respond because of

the uncomfortableness of the social situation, or as a protective device

against middle class meddling, should not be interpreted as meaning that

the child lives in a verbally destitute environment,or even that the

child does not emphasize the importance of verbal manipulation (See

Labov 1969).For example, Shay, Fasold and Wolfram recently conducted

interviews with 45 Puerto Rican and Negro boys from Harlem, ranging

in age from 13-17. The school reoords of the boys in English would no

doubt-indicate that their writing and oral expressionare far below

the middle oleos standard.But consider their responses on a sentence

completion drill designed to get at certain indigenous cultural values.

As part of this drill they were asked to complete the sentence with what

they considered the most appropriate reply-- not in terms school ex-

pectation, but with a culturally appropriate solution. Thus, given

the sentence "el...,t__..L2Laitli_jtjaitherlsoutIfouwannabwita

the vast majority of the respondents replied withouthesitationem Botta

rap to her, or ahze....iveamodrzaoutt. genks, in black ghetto

culture, refers to a distinctively fluent and livelyway of talking,

characterized by a high degree of personal style (SeeKochman 1968:27).

Linguists therefore assume that the label "verbal destitution" cannot

refer to vernacular language patterns, but only to non-indigenous social situations which create suchan expression.

Some linguists, following Chomsky (1965), wouldassume the communicative adequacy of any language or language variety on the basis of an innate "universal" grammar (i.e. it isa putative attribute of being human). This innate language propensity involves the following - 8 -

properties, according to Chomsky (1965:30):

(i) a technique for representing input signals

(ii) a way of representing structural information about these signals

(iii) some initial delimitation of a class of possible hypotheses about language stru'ture

(iv) a method for determining what each hypothesis implies with respect to each sentence

(v) a method for selecting one of the (presumably infinitely many) hypotheses thatare allowed by (iii) and are com- patible with the given primary linguisticdata.

Other linguists, following the behavioralisttradition explicated

by Skinner (1957),insist that theacquisition of language must be

attributed to a stimulus-response relationshiprather than an innate

universal grammar.From the perspective, the adequacyof language

systems would be assumedon the basis of cross-cultural comparisons.

That is, the postulate about the communicativeadequacy of languages

is derived inductively, based on the empirical data from a represen-

tative sample of world languages. Both approaches, then, would make

the same claim about the adequacy oflanguage systems.Although both

approaches arrive at thesame conclusion with respect to this issue,

there is one important implication whichshould be brought out. Chomsky's

perspective assumes thatany normal child will have the equipment to

deal with the logical operationsunderlying language--it isan attribute

of the human mind (See ahomoky 1968).But it is possible, given the behavioralist perspective, thata particular type of environment might inhibit the acquisition of these logicalproperties necessary foran adequate language system.

The question of adequacy ofnonstandard dialects asa communicative systems brings outa very important matter on how one viewsa nonstandard - 9 -

langmage variety. In actuality, it is much broader than thelinguistic

situation, reverting back to the basic approach todifferent social

groups. One can, for example, view black ghetto culture and language

in terms of two basic models, which Beasts (1960) hascalled a deficit 7 model or a difference model. A deficit model treats speech differences

in terms of a norm and deviation from thatnorm, the norm being middle

class white behavior. From a sociological perspective, thismeans

that much of black ghetto behavior, suchas matrifooal homes, is viewed as a pathology. In terms of speech behavior, Black English is considered, in the words of Hurst, (1965:2) "the pathology of non-organicspeech deficiencies". On the other hand, a difference model, whichseems to be much more common to anthropology than sociology andpsychology, considers socially subordinate societies and language varietiesas self- contained systems, inherently neither deficientnor superior.

Although this dichotomy between a deficit and differencemodel may be somewhat oversimplified, it setsa helpful framework for theoretical approaches to nonstandard dialects.But there is also a practical importance for such a distinction. If, for example, one simply considers nonstandard dialects to be corrupt approximationsof standard

English, one may miss important structural facts about thenature of these dialects. For example, consider the following interpretation of the finite use of the form be, a commonly cited feature ofBlack English.

Ruth Golden, who views Black English in terms ofa descending scale of

The different models for describing nonstandarddialects were originally explicated by Cazden (1966). -10-

deviation from standard sh states:

Individuals use different levels of language for

different situations. These levels vary from the

illiterate to the formal and literary. For

instance, starting with the illiterate, Be don't

be here, we might progress to the colloquial, He

ain't here, to the general and informal He isn't

here up to the formal and literary, He is not

loggi. (1963:173).

From the perspective of a deficit model, 11, is simply considered a

corrupt approximation of standard English. The possibility that be may have a grammatically different function is precluded.Instead, it is only considered as a "substitution" for the finite forms of standard English am, is and are.The linguist however, looks at this use of be descriptively; that is, he asks what the grammatical function of this form is regardless of its social consequences. When suchan approach La taken, we find that the form be representsa grammatical category which seems to be unique to:Black:English. This, of course, is not to say that all linguists will accept a given descriptive analysis of this form (See Wolfram 1969:188-196) althougha number of analyses agree that it is used to represent an habitual action of some type.

This type of disagreement is no more serious than the disagreements that linguists may have over the function of the have auxiliary in standard

English.Common to each description of be, however, is the rigorous method of linguistic analysis and the assumption that this form has a linguistic function in its own right. The insistence of language varieties as systems in their own right (with both similarities and - 11 -

differences to related varieties) is the reason that linguists look with suspicion when they see such terms as "substitutions ", "re-

placements", "omissions", "deviations", etc.Such terms used with

reference to nonstandard language varieties imply a value judgement

a.. -,:t a given variety's relation to the standard variety.Terms like

"correspondence" and "alternation" do not have these same implications --

they are statements of fact about language relations. While the termi-

nology may seem to be a trivial matter for the linguist to pick on, the

association of such terms with the deficit type of approach raises a

danger signal to the linguist. To take the position that nonstandard

constructions are simply inaccurate and unworthy approximations of

standard English can only lead to an inaccurate descriptions of what

is assumed to be a self-contained system, which is perfectlyadequate 8 for communication.

Our previous point concerning the adequacy of nonstandard varieties

of English as a system of communication naturally leads us to our next

premise concerning language, namely, that it is systematic and ordered.

Any view of language differences which treats them as unsystematic and

irregular will thus be categorically rejected by the linguist.It is

In terms of sociolinguistic situations, it is quite common fora socially dominant culture to view a socially subordinate one as having an inadequate means of communication.This view is a common manifestation of linguisticethnocentrism of the dominant classes.Thus, Spanish-speaking South Americans often consider the Indian peasants to have no valid language systemverbally destitute.The current treatment of nonstandard English varieties no different, although it may be more subtlebecause Americans have sometimes denied the sociological facts concerning the subordinate role of some segments of the population in American society. - -

assumed that descriptive data of related languages will always reveal

regular and systematic correspondences between different types con-

structions.One can readily see, then, why the linguist reacts

negatively to a view of nonstandard language as that offerred by Hurst, who subsumes differences between Black English and standard English

under the rubric "dialectolalia":

...dialectolalia involves such specific oral abber-

ations as phonemic and subphonemic replacements,

segmental phonemes, phonetic distortions, defective

syntax, misarticulations, mispronunciations) limited

or poor vocabulary, and faulty phonology.These

variables exist most commonly in unsystematic, multi-

farious combinations. (1965:2)

The above position unambiguously treats Black English as an irregular, unsystematic and faulty rather than a different but equal

system.Furthermore, such a position can only be taken when actual

descriptive and sociolinguistic facts are ignored, for the linguist would claim that all evidence points to differences between standard

Ehglish and Black English which are systematic and regular. Take, for example) the case of word-final consonant clusters in such words as test, Emma, and cold.In Black English, the final consonant is regularly absent, the result of a systematic correspondence of a single consonant in Black English where a cluster is found in standard

English. Thus, we get tes', Brouns, and col' in Black English. But these final consonants are not absent randomly or unsystematically.

We observe that the correspondence of a single consonant for a word- final cluster only occurs when both members of a potential cluster are - 13 -

either voioed or voiceless, such as et, Egi, ek, and ld. But when one of the members is voiced and the other voiceless, as in the clusters

(as), lt (2211) and nt count this correspondence does not occur. Instead, Black English is like standard Eoglieh in that both members of the cluster are present. The view that differences between related language varieties are random and haphazard is dangerous not only because it conflicts with a linguistic assumption but also from a practical viewpoint. It can lead to an unsystematic approach in teaching standard English and the teaching of points that may be irk- relevant in terms of the systematio differences between the two language varieties.

As a final premise of the linguist, we must observe that language is learned in the context of the oommunity.Linguists generally agree that children have a fairly complete language system by the age of 5 or 6, with minor adjustments in language competenoe occurring sometimes until 8 or 9. This system is acquired from contact with individuals in their environment. Whether this is primarily the parent - child relationship (which same claim for the middle class white community) or from child peers (which is sometimes claimed for the black ghetto community) their language is acquired through verbal interactionwith individuals in the immediate contest. Whether one maintains that the child has the innate capacity to search for abstract grammatical rules from which sentences are generated (a la Chomdky)or one insists on a behavioralist perspeotive ( a la Skinner), it is presumed that the child will have established an overall language competence by theage of 4-6. The rate of development is .generally assumed to be parallel for children of different social groups (seeSlobin 1967 for an actual - 14 --

investigation of this question),lower class children learning the

nonstandard dialect atapproximately the same rateas middle class

children learning thestandard variety of English.This assumption of the linguist concerning the rate of language developmentagain

comes into basio conflict withbogie statements of educational

psychologists suchas Engelmann, Bereiter and Deutsch,who speak of the communal "language retardation" of ghetto children. Bereiter concludes:

By the time they are five years old, disadvantagedchildren

of almostevery kind are typicallyone of two years retarded

in language development.This is supported by virtuallyany

index of languageone cares to look at (1965: 196)

Any linguist willlook at sucha conclusion with immediate

suspicion. Closer investigationof this claim reveals thatthe fact that these children do not speak standardEnglish is taken tomean that they are linguistically retarded, and, inmany cases, that they

cognitively deficient. Thus, if a black lower classdhild. says HA nice,

a correspondence of the presenttense standard English Re'snine, it is considered to be =underdeveloped standard Englishapproximation

and equivalent to theabsence of oopula ata particular stage of

standard Ebglish development (See,for example, Bereiterand Engelmann

1966: 139-140). The fact that this formis used by adult speakers

is irrelvant, only meaningthat adultsmay have some stabalized form

of language retardation. The linguist, however,suggests that Black English is simply one of many languages and dialects,including Russian, which have a zero copula realization inthe present tense. No meaning is lost; an "identity statement" is justas permissible in this dialect as any other languageor dialect. This form hapno relation - 15 - to the ability or inability to' conceptualize. Similarly,-Aaitoiy-dis- crimination tests (such as Wepman's 1958) which are designed on a standard English norm are de-facto dismissed by the linguist as biased against the nonstandard system. The learning of standard English must be clearly differentiated from language development of an indigenous dialect. Careful attention should be made, from the viewpoint of linguistic relativism, in order not to erroneously transfer legitimate dialect differences into matters of language acquisition.

The linguist, in support of the linguistic equality of nonstandard dialects, considers evidence on relative language proficiency as that recently provided by Baratz (1969) to be an empirical justification for his claims. Baratz conducted a bidialectal test in which she has oompared the proficiency of a group of black ghetto children in re- peating standard English and Black English sentences.As might be expected, the bleier. children were considerably more proficient in re- peating the Black English sentences. When they repeated the standard

English sentences, however, there were predictable differences in their repetitIons based on interference from Black English. The same test was Vaen administered to a group of white middle class suburban children, who repeated the standard English sentences quite adequately, but had predictable differences in their repetition of the Black English sentences based on interference from standard Ihglish. Which of these groups, then, was linguistically retarded? We must be careful not to confuse social acceptability, and no one would deny the social stigmatization of nonstandard dialects, with language acquisition.

In sum, the relativistic viewpoint of the linguist emphasizes the fully systematic, but different nature of nonstandard dialects. This -16-

committment, however, should not be taken to mean that linguists are necessarily for the perpetuation of nonstandard dialects in the face

of middle class intolerance. We are certainly not so naive as to

suggest that standard English is not a prerequisite for making it in middle class society, and the person who desires to do so must be

given that option.The stigmatization of nonstandard dialects, even by those who regularly use them as their only means of communication, is such that learning standard English may be a socially expedient

correlate of upward mobility. But these social issues, the linguist argues, have nothing to do with the inherent capacity of nonstandard dialects for effective communication within their prescribed social .17-

II Current Research

Within the discipline of linguistics, it is the field of dialec-

tology which was responsible for the earliest attempts to account for

social variation in speech. American dialectologists recognized that

social differences had to be considered, even though the primary goal

of dialect geography was the correlation of settlement history with

regional varieties of English. Kurath, for example, in directing the

Icnguistic Atlas of the United States and Canada was aware that social

differences intersected with settlement history and geographical

differences to account for linguistic variation.As reported in the

Handbook of the Linguistic Geography of New England, Linguistic Atlas

fieldworkers divided informants into three main types, as follows:

Type I: Little formal education, little reading and restricted social contacts.

TypeII: Better formal education (usually high school) and/or wider reading and social contacts.

Type III: Superior education (usually college), cultured background, wide reading and/or extensive social contacts (Kurath 1939: 44).

In addition, each of the above types was subdivided as:

Type A 'Aged, and/or regarded by the fieldworker as old fashioned.

Type B Middle-aged or younger, and/or regarded by the fieldworker as more modern (Kurath 1939: 44).

Although different social types were recognized in the work of tee

Linguistic Atlas, several difficulties were apparent because the social parameter was not adequately considered. The social classification of informants was dependent on the fieldworkers' subjective impressions.

The vagueness with which the social types were profiled (e.g. "little - 18 - readtag and restricted social contacts") caused the social classification of informants to be unreliable. Furthermore, no verifiable sociological model for rating the social status of informants was utilized. Education, which seemed to be primary in the evaluation of informants, is only one of the various factors which is used by social scientists in rating social status.

Whereas the correlation of social with linguistic differences was of secondary concern in the work of the Linguistic Atlas (Kurath 1941,

1949), later interpretation of the Linguistic Atlas data gave more direct attention to the importance of social factors in accounting for linguistic diversity. Dialectologists, however, still seemed to appeal to the social parameter only when "data proved too complicated to be explained by merely a geographical statement or a statement of settlement history" (MoDsvid 1948: 194).Thus, MoDavid's "Postvocalic

-r in South Carolina: A social analysis"(1948) amends a geographical explanation of postvocalic -r in the Piedmont area of South Carolina by analyzing the intersection of social class with geographical differences.

As will be seen later, dialectologists continue to work with the social consequences of speech variation, but the methods of "mainstream dialectologista" such as Kurath, McDavid, and Pederson have actually changed very little.

From another perspective, anthropological linguists have made significant contributions to the study of linguistic correlates of social stratification in the last decade.Whereas dialectologists have been satisfied with rough approximations of social divisions to which linguistic phenomena may be related, anthropologists have characteris- tically been rigorous in their differentiation of social groups to which 19 linguistic variables may be related. Independent ethnographical der. scription of behavioral patterns characterizing different social strata is required before any correlation of linguistic variables with these strata can be made.Research on the social stratification of linguistic features has been pioneered by Gumperz (1958a, 1958b,

1961, 1964), Hymes (1961, 1964), and Bright (1960, 1964, 1966). For example, Gumperz, in several articles (1958a, 1958b), has shown how linguistic variables, particularly phonological variables, relate to the, caste systems of India.Southeastern Asia, perhaps because of its rigid stratification between castes, has received the most extensive consideration by anthropological linguists.Anthropological linguists such as Hymea and Gumperz have concerned themselves with developing a structural taxonomy of the factors which must be dealt with from a sociolinguistic perspective of verbal behavior, such as settings, participants, topics, and functions of interaction. Little consideration has been given to American English by anthropological linguists, although

Fisher (1958) provided an analysis of the morphemic variation between the suffixal participle /42)/ and /4En/ in English by considering the social background of 24 children in a New England town.

It was Labov's work on the social stratification of English in New

York city (1963a, 1963b0 1.965a, 1960, 19660), more than any other research, that has sharpened the theoretical and methodological bases for sociolinguistic research.Using a survey by the Mobilization for

Youth as his sociological model, he analyzed the speech of over a hundred, randomly selected informants.Five different phonological variables

(ph, eh, r, th, dh), isolated in four contextual styles (careful speech, casual speech, reading, word lists) were correlated with the social - 20 -

stratification of the informants. Idbav made several major contributions

to the study of linguisticcorrelates of social stratification. In the

first place, he used sociologicallyvalid procedures in selecting the

informants for his sample. Many linguists prior to Labov were largely

satisfied with biased, non-randominformant selection. Also, Labov's

quantitative measurement of linguisticvariables, although not the

first, was considerablymore extensive than any previous sociolinguistic

research.Further, his effort to isolate contextual styleson the basis

of extra- linguistic "channelcues" was a careful attempt to define

interview styles in linguistics.The major contribution of Labov was

his demonstration that speechdifferences within a community, often

dismissed by linguistsas "free variation", systematically correlated

with social differences:

The Detroit Dialect Study (ShunWolfram and Riley 1967), exper-

imented with several different methodsof analyzing speech differences.

It extended the insights of Labovon the linguistic variable to grammatical as well as phonologicalvariables. An attempt to measure differences by the quantitativemeasurement of structural types (e.g. clause and phrase types)was also investigated.

Despite a developing sociolinguistictradition within linguistics over the past several decades, the actual structuraldescription of non- 9 standard dialects has received littleattention.

Nonstandard dialects have, ofcourse, always received incidental attention in prescriptive Ehglidh textbookswhich point out "in- correct" speech patterns to be avoided. - 21 -

To a certain degree, this lack of attentioncan be attributed to the attitude that nonstandard speech is less worthyof interest than the study of socially acceptable varieties of lish. Another con- tributing factor for this neglectmay have been the assumptionthat the nonstandard dialects were minimally different intheir structure and that when comprehensive studies of standardEnglish were completed, it would be a relatively simple matter to adjustgrammatical dei soriptions to include nonstandard varieties.With respect to Black English, descriptive attention wasno doubt delayed by dialeotologists who maintained that it was not essentiallydifferent from the speech of Southern whites of comparablesocio-economio levels.As an example of such a view, note KUrathisconclusions from his workon the Iasuistio Atlas:

By and large the Southern Negro speaksthe language of the white man of his localityor area and of his

education.... As far as the speechof the uneducated

Negroes is concerned, it differs little fromthat of the illiterate white; that is, itexhibits the same

regional end localvariations as that ofthe simple

white folk. (1949:6)

Stewart (1965:13) also observes that thestructural neglect ofBlack English may also have been related toconcern for the feelings of Negroes:

As this Fthe studyof Black Eh:els:111 relates to the

speech of Negroes,it has been reinforced bya

commendable desire toemphasize the potentialof the Negro to be identical with white Americansand ao- - 22

cordingly to deenphasize any current behavioral pat-

terns which might not seem to contribute direotly

to that goal... respect for the feeling of Negrnes

themselves has probably played a part in discouraging

the study of Negro speech.For, as is quite under-

standable, many Negroes (particularly educated ones)

are somewhat sensitive about any public focus on

distinctively Negro behavior, particularly if it

happens to be that of lower class Negroes.

Whatever the reasons may have been, it was not until the last few years that the study of Black English was seriously undertaken.

Although there are several current research projects on the linguistic structure of Black English, by comparison, there are still only limited number of linguists who have taken a serious interest in this area.

Stewart (1964, 1967; 1968) and Bailey (1965) probably did more to turn the attention of linguists to the study of Black English than any one else, partly because their work chronologically preceded other linguists and partly because of their dogmatio rejection of the dialectological treatment of ethnic differences in speech.Coming from oreolist backgrounds, both Bailey and Stewart maintained that

Black English was not identical to the speech of Southern whines of a comparable socio-eoonomin class, but significantly different. Bailey, for example, noted:

I would like to suggest that the Southern Negro

"dialect" differs from other Southern speech

because its deep structure is different, having -2. -

its origins as it undoubtedly does in some Proto-

Creole grammatical structure. (1965:172)

Obviously, such a position comes into sharp conflict with the

traditional position suggested by a number of American dialectolo-

gists. What then, can account for this Sharp difference of opinion?

One explanation is that dialectologists have focused their attention

on the similarities between nonstandard Negro dialects and white

dialects, whereas creolists have focusedon the differences between these two varieties of English.Dialectologists have been largely oocupied with phonological and lexical differences, the levelson which the dialects are nearly (but not completely) alike.Creolists, on the other band, have concerned themselves with subtle differences between grammatical categories. Stewart has mainly concentrated on the historioal relations of Black English towhat be considers a creole origin.He notes:

Of those Africans who fell victim to the Atlantic slave

trade and were brought to the New World,many found it

necessary to learn some kind of English. With very few

exceptions, the form of English which they acquiredwas

a pidginized one, and this kind of English became so

well-established as the principal medium of communica-

tion between Negro slaves in the British coloniesthat

it was passed on as a creole language to succeeding

generations of the New World Negroes, from whom itwas

their native tongue (1967: 22).

. Present day Negro dialect, according to Stewart,has resulted from a process which labels "decreolization" (i.e.the loos of creole - 24 - features). Through contact with the British - derived dialects the creole variety of Englieh spoken by Negroes merged with other dialects on thglish.The merging process, howevefc was neither instantaneoua nor complete. Stewart assertii:-

Indeed, the non-standard speech of present-day

American Negroes still seems to exhibit

structural traces of a creole 1.740ecessor, and

this is probably a reason why it is in some ways

more deviant from standard English than is the

non - standard speech of even the most uneducated

American whites (1968: 3).

Stewart substantiates his claim that Negro dialects are derived f r o m a w i d e s p r e a d s l a v e c r e o le b y e xamining the close relationship which is found between 18th and 19th century Negro dialect and other

New World creoles (Stewart 1967). His source for the study of 18th and 19th century Negro dialect is the representations of Negro dialect used in the literary works of this period. Although this may seem like an unreliable source, Stewart's exhaustive knowledge of the literary records of Negro dialect during this period and his apparent ability to evaluate the reliability of the various authors makes his historical documentation quiteplausible./0Furthermore, Stewart's familiarity with a number of different creoles, including Gullah and the Caribbean °reales, causes one to consider seriously any statement he makes concerning the relations of various Black English structures

10. It would, of course, be nice if Stewart would put his actual method for determining the reliability of the various literary records in print, so that this type of information would be avai.:4ible to the general public. m 25 -

to a creole predecessor.Although Stewart's thorough knowledge of

Black English certainly can not be disputed, severalpoints he makes

do not appear to be as clear-cutas he asserts. For one, his approach to analysis concentrateson particular items rather than a wholistic approach to the structure of Black English.An attempt

to assemble a comprehensive inventory of differences betweenSouthern white and Negro of comparable socio-economic classesmay lead one to a considerable smaller list than anticipated.

Furthermore, the origins ofsome of the items would certainly be disputed by dialectologists.Others night be disputed on empirical grounds.For example, Stewart observes that implosive stops, which he claims are quite easy for the trained phoneticianto perceive are unique to the Black Ebgliah speaker. But there are some linguists who would claim that the American English stopcan sometimes be implosive.

Furthermore, I know of three reasonably competent phoneticianswho agree that both Black and white speakers use implosives. At any rate, the issue is not nearlyas clear-cut as Stewart makes it out to be.

Finally, Stewart emphasizes differences between Black English and standard English as opposed to similarities. This in itself may be justified since it is the differenceswhich cause interference problems between dialects. It must be pointed out however, that the inventory of differences is much smallerthan the inventory of similarities.In addition, the clear majority of differencesseem to be on a surface rather than an underlying level (See, e.g. liabov, et al

1968). In most inventories of differences between BlackEbglidh and other English varieties, the listsare quite restricted.An expansive list is lacking, either because the listis simply not as exhaustive as -26-- suggested or because descriptive data are still lacking.

From a purely, descriptive viewpoint there are several current projects which merit attention. Probably the most radical of these is offered by Loflin (1967.1 1968) formerly of the Center for Applied

Linguistics and presently of the Center for Research in Social Behavior at the University of Missouri. Loflin considers the differences between standard English and Black English to be of such significance that

Black Englidh be treated as a foreign language. Be observes:

Efforts to construct a grammar for Nonstandard Negro

English suggest that the similarities between it and

Standard English are superficial. There is every

reason, at this stage of research to believe that a

fuller description of Nonstandard Negro English will

show a grammatical system which must be treated as a

foreign language (1967?: 1312).

In justification of his treatment of Black English as a significantly different system (i.e. different in its underlying structure) from standard Enefish, he has described the verbsystem.11Be concludes that aspect dominates over tense in Black English, whereas the opposite is true for standard English. A careful look at his description reveals that it must be challenged both on empirical and theoretical grounds.For example, one of the basic justifications for his description of the verb system of Black English is the absence of the awdliary have in the running test of his eirelinformant;

11. This choice is by no means accidental since most linguists agree that if there are any significant differences between Black English and Standard English, they will be found in the verb system. - 27 -

empirical investigation of the staff at the Center for Applied Lin-

guistics in Washington, Labov in New York, and Wolfram in Detroit

clearly reveal the underlying presence of have (although itmay be

deleted by a low level phonological rule). In fact, one of the

striking things about Black Englishseems to be the extraordinary

use of the past form of this construction in narrative discourse

(e.g. Be hadcame to the store).

Other parts of Loflin's analysis of the verbal system reveal

a neglect of the overall pattering of Black English. Thus, for

example, the clear evidence thata phonological pattern isrespon-

sible for the absence of most past tense -edsuffixes is overlooked

(see e.g. Wolfram 1969:71-74).Aside from these types of important

oversights, Loflin shows some inadequacies in theuse of the model

in which he describes the structure of BlackEnglishtransformational-

generative grammar.Even if one assumes that Loflin's analysis of

the passive is correct (Loflin 1968), Loflin's passive ruleis

entirely unworkable in any current theory of generativegrammar.

(i.e. it does not generate whatLoflin says it generates).Although

Loflin's work certainly showsa high degree of creativity, his general

approach and specific description of the Black Englishverb system

can hardly be considered valid.

A somewhat different attempt to describe the linguisticstructure of Child Black English in Florida is offered byHouston (1969).A number of phonological "rules"are given, but no grammatical rules since, according to Houston,"only four majorsyntactic differences between Child Bleak and standard White Englishhave appeared" (1969:

606). To those linguists seriously attempting todescribe the structure -26-

of Black English, Houston's description shows significant theoretical

and empirical flaws.From a theoretical standpoint, her approach to

the description of Black English cannot be conaidered acceptable ro

from any current taxonomic or generative standpoint. Her rules are,

by her own admission, nothing more than a set of correspondences which

relate Child Black English to White Standard English, yet she sets up

her correspondences in the form of processes so that they have the

form of pseudo rewrite rules. Her justification for this curious

device is "convenience", hardly a sufficient reason for the theoretical

.or descriptive linguist. In essence, what she does is derive surface

forms in Black English from surface forms in standard White English.

The rules are even given as ordered, yetany descriptive linguist can

see that they are not ordered in the sense that this concept is used

from any standpoint in linguistics.

Some of the rules she gives also lack formal motivation. Although

she mentions general postulates which govern the treatment of phenomena

as phonological instead of grammatical ("their relative generality in

the language as a whole, and the importance of the grammatical claims"

(1969:603),some of the rules she treats as phonological can be

seriously disputed.Why, for example, is the third person singular

-Z a phonological rule rather than a grammatical rule? Thirdperson

singular -Z affeots all verbs, not only those involving consonant

clusters (e.g. it affects boos as well as dreams).Yet, lack of

formal motivation for the correspondence is lackingso that the rules

appear to simply be ad hoc.Some of the rules which are given,

furthermore, do not describe the data which they presumablyare supposed

to account for. Thus, as the rule for consonant cluster reduction is -.29 -

formally stated, itcan account only for bimorphemio clusters (e.g.

it can account forguessed being realizedas /g a/, but not a

monomorphemic cluster libeigatt,which is realized as /gta4. Such

apparent oversightsare, unfortunately, characteristic of the rules.

And finally,some of the empirical data she displaysare suspect.

Stewart, Labov, and thestaff at the Center for AppliedLinguistics

working with Black Ebgliahwould all have critical disputes about

some of her observed data. EVen if she treatsa number of apparently

grammatical phenomena as phonological, she does not mentiondifferences

in verbal paradigms,modals, person agreement,existential it, ple-

onastic forms other thanpronominal apposition, etc. In the light of

these theoretical andempirical inadequacies, Houston'sstudy cannot be considered as a serious attempt to describe thestructure of Black

&gilt& froma linguistic perspective.

The research of Labovand associates (1965, 1960on the structure of the nonstandard speech of Negroes and PuertoRicans in New York is the single most exhaustivestudy of a nonstandard speech community bmailable.12 Having already ol.boa oontTibutionsthat Labov made to sooiolinsuiation in his study of the overall populationof New York City, we must again cite a number ofsignificant sociolinguistic innovations in his Harlem research.In the first place* hisstudy of language in the settingof an adolescentpeer group broke with the more traditional Individualinterview we thod of 000i of insui Furthermore, he has described both the functional and structuralanimerta of the nonstandard vernacular.The structural description of Black

Ebel& included more featuresof the phonology andgrammar of Black

Ebglidh in detail thanany other single description. In addition to

12. Although Labov includesPuerto Rican speech in his title,his actual description is limitedto the Negro population. . 30 . have on variability. This is not to say that a statement of the relevant environments in which so-called "free variation" took place was not a requisite for adequate linguistic description.

But the recognition that certain environments may affect the occurrence of a given variant much more than others was character- istioally absent. Yet, the variables described by Labov and others

(See, e.g. Wolfram 1969) show. that certain types of linguistic environments intersect with extra-linguistic factors to account for variation between forma. Labov, therefore, has suggested that the* notion of linguistic and non-linguistic constraints be incorporated into the formal representation of a linguistic rule. He has thus proposed what he calls the variable rule (1968:24). By introducing the variable rule, Labov attempts to formally incorporate the constraints (linguistic and non-linguistic) which directly affect the variability of items.To achieve this end, Labov suggests that "we associate with each variable rule a specific quantity which denotes the proportion of cases in which the rule applies as part of the rule structure itself" (Labov 1968:25). The value of a variable rule is defined as a function of the constraints which limit the categorical operation of the rule. This may be represented as:

f = 1 - (a + b + c + n) where f= the frequency of application, 1 the categorical operation of a rule, and a, b, c, ...n the various constraints limiting categorical rule application (i.e. the variable input). The constraints are "ranked" -- ranked in sense that certain linguistic environments clearly outweigh others in their effect on variability (e.g. LW)? c, ...n). 31...

Labov's creative innovations in field methods and his comprehensive

poninlingnintin docsoriptdon of Black English, he has carefully

examined the implications that his researchhas for theoretical

linguistics.Based on his elicitation ofpeer group speech in a

relatively spontaneous setting, he has observed thatmany of the

variants associated with Black EnglIshmust be considered "inherently

variable" with more standard -like variants. That is, fluctuation

between many variants seem to be inherentto the vernacular structure 13 and not simply an "importation" froma superimposed variety.

Evidence for this is found in the systematicways in which certain

types of fluctuation seem to operatewithin the most indigenous speech

situation./4Lam points out that independent linguisticas well

as social variables must be considered in describing thesystematic

variation of forms. The correlation of sociological with linguistic

variables to account for fluctuation betweenforms has become well-

established within the last decade ofsociolinguistic research. But

the notion of systematic variationas a function of independent lin-

guistic variables has not been consideredseriously. The fact that

linguistic environmentcan greatly affect the variability of items has some important implicationson the concept of "optionality" in

linguistics. The limitation of linguistics to qualitative,discrete units has somehow precludedany affect that linguistic environment may

13. This position does not precludethe possibility that historically, alternations may have been importations. It simply means that from a synchronic standpoint, fluctuationis an inherent part of the Black English system.

14. For a more detailed discussion of the differentiationof "dialect mixture" from inherent variability,see Wolfram (1968:43-47). ..32

Labav's careful .72amination of theBlack English system and field techniques leave little room for criticism.The description of certain features will certainly not find unanimousagreement (See, e.g. Wolfram

196: 192-194), but Labav's overall study is a most importantsociolin- guistic description.His incorporation of the variablerule into a formal grammar will, however, stir considerabledisagreement among linguists. The controversy over the rule does not oonernits obser- vational adequacy, but whether this can and shouldbe included in the formal description of a grammar. Is this rule simply part of a

"performance" model, and, as such, irrelevant to thedescriptive adequacy of a grammar, or is this an integral part of language"competence"? The quantitative figures which can be assignedto various constraints would seem to be part of performance model,but the regular and hierarohical effect of various linguisticoonstraints on variability cannot be dismissed quite as readily. This is, no doubt, as issue that is destined to be ofconsiderable importance for theoretical linguistics.

The research undertaken by the SociolinguisticsProgram at the

Center for Applied Linguistics deals both withthe linguistic correlates of stratification in the Negro community and thestructural description of Black English. Data from several different locations is being analysed, including Washington, D.C., Detroit, Michigan(a continuation

of the Detroit Dialect Study under the direction of RogerShuy), and more recently, Lexington, Mississippi.Wolfram's study of the Detroit

Negro population demonstrates how several classesof Negroes are

differential on the basis of grammatical and phonological variables.

The role of social status, sex, age, and racial isolation areall shown 33

to correlate with linguistic differences. In addition, the extent to

which the social differentiation between linguistic variables is

quantitative or qualitative, the relation between social diagnostic

phonological and grammatical variables and the effect of independent

linguistic contrasts in variability are examined.

The investigation of phonological and grammatical variables reveals

that the phonological differences between social groups tend to be

quantitative whereas the grammatical 4ifferenoes are often qualitative.

Three of the four phonological variables (word -final consonant clusters,

syllable-final d, and postvocalic r) indicate that the social groups are

differentiated primarily on the basis of the relative frequency of variants. Only the 0 variable, which shows the categorical absenoe of the

f variant in middle -class speech, indicates a qualitative difference between social groups. On the other hand, all four grammatical variables (multiple negation, suffixal -Z, copula absence, and in- variant be) reveal the categorical absence of certain variants among middle-class informants.

By introducing the concepts of "sharp" (i.e. a significant difference in the frequency of particular variants between contiguous social groups) and "gradient" (i.e. a progrescive difference in the frequency of particular variants between social groups) an important difference in the way phonological and grammatical variables stratify the population can be observed. Grammatical variables usually show sharp stratification, whereas phonological variables show gradient stratification. All the grammatical variables investigated in the study reveal sharp stratification, whereas three of the four phonological variables indicate gradient stratification. T- 34 -

Finally, Wolfram's research demonstrates that it is impossible to arrive at an adequate understanding of the nature of sociolinguistic variation without considering the effects of independent linguistic constrants. In accounting for frequency differences among variants it is essential to consider the effect of linguistic environment.

Wolfram's work reinforceirmany of the conclusioas that Iabov has independently come to in his research in New York, suggesting that there is considerable uniformity in the sociolinguistic patterning of

Black Fnglish in large, Northern metropolitan areas.Wolfram's limited sample (48 informants), however, needs further extention, particularly with reference to his conclusions about age, sex, and racial isolation. Statistical sophistication is also lacking in some of his conclusions based or quantitative differences.Finally, the functional reasons for certain types of differences, although important, are not examined.For example, is the pattern of sex differentiation due to different types of contact situations that males and females have with the socially superordinate white community

(e.g. female domestics working in close contact with middle-class white females) or is this an indigelJous behavioral characteristic of the Negro community (e.g. the use of socially stigmatized forms is a symbol of masculinity)?

Fhsold's research in Washington, D.C. currently includes a study of the social stratification of speech in the Negro community and the structural description of various features of Black Engliah.Recently

Fasold (1969) has explicated one of the crucial issues for sociolinguistic analysis and the representation of sociolinguistic information from a linguistic perspective, namely, "implicational analysis", - 35:

"frequency analysis", or a combination of the two. The former approach deals with the implication of the presence of certainsocially diagnostic linguistic features for thepresence/absence of others; frequency analysis involves the variabilityof linguistic features asthey relate tosocial class, and the combination ofthe two approaches used the statistical method of "factoranalysis" to deal both withthe frequency of occurrence and the co-occurrencere- strictions of variants. In investigatingthese various approaches,

Paso ld suggests that the more adequate approachis probably the one that can most readily incorporate the insights ofthe other. He concludes that frequency analysis can incorporatethe insights of continuum analysis by simply including an "invariancecategory," whereas continuum analysis must arbitrary assign anyobserved variability between features into a binarycategories. He submits that the third approach, that of combiningcontinuum and frequency approaches viafactor analysis is the leastrevealing because it only leads to groupings already obvious.Also, there is noapparent way to ineoxperratofactor analysisinto linguistic) theory.It appears, however, thatFasold has dismissed the thirdalternativetoo lightly.

Theoretically, it holds thepotential to reveal less thanobvious continuum sets, and to validateapparent groupings.If it then proves to be valid, it is the tank of linguistio theory to incorporatethis concept. Another important area currently beinginvestigated by Roger Shuy and colleagues is-that of relation between white southern and Negro speech of comparable socio-economic classes, based on data from Lexington, Mississippi.Although still at a prelim' inary stage 3.

of analysis, it is hoped that these data will reveal concrete answers

to the controversy of Negro/White speech differences in the deep

South.

Although the above mentioned projects describe current major

research dealing with the study of nonstandard dialects in the United

States, there are several other studies that can be mentioned briefly.

We have already seen how the work of American dialectologistswas one

of the earliest attempts to deal with social factors in linguistic

diversity within linguistics.The more recent work of Naomi& and

Austin (1967), Pederson (1965) and Williamson (1961) indicatesa

continued interest in this area.However, retention of Linguistic

Atlas techniques, now superseded bymore sophisticated sociological

and anthropological techniques, places such research ata serious

disadvantage. The continued emphasis on lexical it an phonology

preclude a comprehensive structural description ofa nonstandard

grammar. Current studies of social dialects by dialectologists have

also neglected the systematic nature of variation that quantitative

studies of variability reveal.Furthermore, the apparent disinterest

in the implications of such research for linguistic theory does not

coincide with the direction of current sociolinguistic studies. :

.There are also several projects which can only be mentioned briefly because of their incipient nature. Fraser and colleagues at the Language Research Foundation are presently beginning the description of Mild Black English in New York City, employing the most up-to-date insights of theoretical linguistic. Fieldwork in this project however, has not yet begun. - 3-7 -

The sociolinguistics aspect (there is also a pedogogical aspect)

of the study by Legum, Williams and associates (1968), under the

auspicies of the Southwest Regional Laboratory at Inglewood is

presently conducting interviews with childpeer groups in Watts. At

this stage, only the statement of the theoretical linguistic and

sociolinguistic foundations of their research is available and these

are derived mainly from Ldbov's research.

The hat Texas Dialect Project, directed by Troike and Galvan

at the University of Texas (1968), has conducted interviews with

over 200 informants in five communities in That Texas, representing

different races, several socio-economic levels, and variousage

groups. The interview involved the elicitation of free conversation

between a fieldworker and two informants. Preliminary exploration

has resulted in the isolation ofa number of different phonological

and grammatical variables for analysis, and the frequencyof socially

significant variants isnow being analyzed.

The Pittsburgh Dialect Project, by Parslow anl colleagues (1968)

is an attempt to identify regional and socialparameters of different

social and ethnic groups, following the modelthat Shuy used in the

Detroit Dialect Study. This project, however, is still involved in

data collection so thatno description of the analysis is yet available.

Finally, a project being carried out by Billiard, Lazarus, and McDavid in Fort Wayne (1969) focuseson the identification of socially diagnostic linguistic features.The prospectus of this project does not include detailed linguisticanalysis, however, and the field procedures are apparently not takingadvantage of the most recent developments in field techniques. -38 " III DIRECTIONS FOR FUTURE RESEARCH

On the basis of our previous discussion of research projects, it should be apparent that some aspects of social dialects are being studied thoroughly while others are neglected. It is therefore the purpose of this section to summarize areas which have been investigated adequately and to suggest the direction that 15 future research might take. These can conveniently be discussed in terms of three main areas: (1) field techniques, (2) descriptive studies, and (3) theoretical issues.

1. Field Techniques

As was seen in the preceding section, sociolinguistic field procedures by linguists have made rapid, progress within the last several years. We now see that the design of fieldwork and sampling procedures can give a reliable representation of the sociolinguistic parameters of a community (See, for example, Iabov 1966%, Chapter 6 or Shuy, Wolfram and Riley 1968, Chapter 2). Current interview procedures have also developed according to social science standards of interviewing (See Labov 1966, Chapter 5 or Shuy, Wolfram and Riley

1968, Chapter 5-7, and Slobin 1967) so that many of the criticisms of

Pickford (1956) concerning the inadequacy of the Linguistic Atlas fieldwork design are no longer applicable to ourrent sociolinguistic research. Furthermore, elicitation procedures, particu]ary as related to stylistic variation, have made significant advances following the insights of Labav (19%01968).

15. These directions, no doubt, reflect the biases of the author. However, many of these directions have been discussed with may colleagues at one time or another, so that they represent more than. personal preferences. jr

There are, however, several areas in which further refinement

of research design can add to the validity and reliability of socio-

linguistic studies.With respect to sampling, we are still not

certain of the most efficient size for a reliable study of social

dialects. What, for example, is the minimal number of informants

in each social "cell" for the linguist to adequately characterize

the linguistic behavior of that cell. It appears that linguistic

behavior is more homogeneous than some other types of behavior

investigated by sociologists so that we can conceivably achieve

reliability using a smaller sample than other types of sociological

surveys.Also, because of the detailed nature of certain types of

linguistic analysis, it is impractical to work with samples the

size of some sociological surveys. But we still do not know what

constitutes a minimally adequate representation for the study of

social dialects. One way of getting at such information would be

to take a reasonably large sample such as the Detroit Dialect Study

(which includedover 700 interviews) and compare several linguistic features using different sizes of subeamples within the large study

to establish a minimal standard fora reliable sample. Information of this type could determine the most efficient size of future social dialect surveys.

Another area where we lack concrete informationconcerns the role of the fieldworker in affecting the speech of informants.We suspect that the race, sex, or social class of the interviewer might be important conditioning factors with respect to speech, and there are several studies which show such factors to correlate with speech variation (e.g. Anahen 1969).But we still need an exhaustive study -40-

of the relative importance of the social characteristicsof the in-

terviewer. For example, is the correlation between therace of the

interviewer and the informant's speechvariation simply a function of

race, or is it actually more related toa person's ability to identify

with the social class of theinformant, or, is it a combination of

these? And, if such correlationsexist, do they affect all socially

diagnostic linguistics variablesor only those on a more conscious

level?These are questions about theinterview which will suggest the

relative importance ofcontrolling interviewer variables.

One area of top priority forfield techniques is the establish-

went of elicitation procedureswhich can get at judgements of the

grammaticality of nonstandard structuresapart from judgements about

social acceptability. The linguist's usual procedure is toobtain a

language sample in order to determinethe rules of the grammar and then

directly ask the native informantwhether or not certain grammatical

contrasts that he reconstructs from hisrules are indeed significant

in his language (i.e.can they be generated by the rules of his grammar?).

This same procedure, however,cannot be used in dealing with the

grammaticality of nonstandardsentences, since it is virtually im-

possible to get such judgementsisolated from social notions of

acceptability (i.e. the MissFiditch notion of "correctness"). Thus, for example, if a linguistwere to aok a Black English informant if a sentence such as 21111IJIOntullmaingwere acceptable, he is liable to have the informant rejectthe sentence.But we cannot be sure if the informant rejected the sentence becauseit is not part of his competence or because ofthe social stigmatization of thesentence.

Ideally, this might beovercome by a linguistically sophisticated -41-

native speaker of Black CL lish. However, in my experience, most

linguistically sophisticated speakers of Black English have also

acquired standard ish, and, in doing so, invariably have lost

sensitivity to the grammatical boundaries of the Black Ehglish vernacular, which are so important in establishing underlying

competence. It is therefore imperative that we develop methods by which we can get at the generative capacity of the Black English grammar rules. In order to do so, we must take advantage of more indirect ways of getting at competence.One important way may be

through the development of different types of "word games".For example, Fasold (personal communication) has been experimenting with a sentence completion technique in which the informant is given a stimulus sentence and asked to respond to the sentence on the basis of a pattern which will determine whether or not the given feature is present in the underlying structure.To illustrate, consider whether or not the underlying auxiliary have is an integral part of the Black English grammatical system. The informant is given a sentence such as They been there Ellamttial, and asked to respond to this sentence by completing the response ;know th2z

If the informant responds by completing the sentence with have,we may be assured that there is an underlying have: however, if he responds by using another auxiliary such as did, then he probably does not have the underlying auxiliary have. The establishment of such indirect techniques to get at competence is important for future structural descriptions of the nonstandard grammatical systems. Of oourse, one must be careful to use stimuli sentences and patterns whichare indigenous to the dialect, which makes fami.darity with the dialect -42- a prerequisite. In developing procedures of this type, linguistic fieldwork can probably profit mostly from elicitation techniques for children (See Slobin, 1967; Menyuk 1969, Chapter 4) at various acquisitional levels, but other new techniques will also have to be established.

Finally, the linguist interested in the social parameters of language is still uncertain about the importance of statistical calculations in comparing the various quantitative measurements that are made. Both

Iabov (1968) and Wolfram'. (1969) rely heavilyon quantitative evidence, but neither uses tests to determine the statistical significance of their quantitative differences.TArguists, because of a tradition of qualitative analysis, tend to ignore statistical calculations. In justification we may say that some of the quantitative differences are so prominent that statistical calculations are hardly needed. In other cases, it is the establishment of the general direction of different frequency scoresthatia more important than the significance between specific figures.Furthermore, the linguist might claim that his dataare far more regular andreproducible than the type of data sociologists are used to analyzing via statistics.

But we may be arguing from naivete. At any rate, the relative importance of statistics for sociolinguistic study is an area which needs careful research and explication. We must know in what areas statistical calculations are expedient, what areas they are question- able and what areas they are inapplicable for the linguist doing research in social dialects. -43-

2. 12eJsr.:zbive Studies

As we noted in our description ofcurrent research projects,

there are several aspects ofsocial dialects in the UnitedStates

which have occupied the attentionof linguists. Corresponding to

the popular focuson Negroes in the inner city,we have witnessed

a number of attempts to describe thegrammatical and phonological

structure of Black English, varying greatly in quality.Research

in New York, Detroit,Washington, D.C., and Wattson the structure

of Black English seems to give adequate representationof this

dialect in the large urbanarea, especially because of the apparent

similarity in the structure of Black English in theseareas. This is not to say that there are no regional differences, but theoverall

structure of the dialectshows striking similaritiesin these different

locations. There are, ofcourse, aspects of the dialect whichhave

not been covered in detail,but the major features ofthe dialect

can be derived by looking at thevarious studies. One descriptive

aspect which has not beencovered in any of thesestudies is intonation, yet most linguists agree that there are substantialdifferences between Black lish and other varietiesof English in thisarea.

The correlation of socialclass with linguisticvariables in large urban dream is also receivingan increasing amount of attention.

The relationship of various parameters ofsocial class haveor are currently being described for New York, Washington,11.C., Chicago,

Pittsburgh, and Port Wayne,so that we are obtaininga representative number of studieson language and social class.

There are, however, still a number of areas which havereceived little or no descriptive attention. In the preceding paragraphone - 44

can note that the majority of studies of Black Englishfocus on

large Northern areas. Ve still need adequate descriptive studies

of Black English in both the rural and urbanSouth. Such studies

must be the first step in comparing the linguisticassimilation

that takes place when mass migrationtakes place, as it did during

the last fifty yearsamong the Negro population. Are Southern and

Northern varieties of Black Englishessentially alike, and, if not,

in what ways do they differ?Only comprehensive studies of the

structure of Black English in selectedareas of the South can

answer this question. Such studies should preferably beselected

to represent differentareas of the South, including the coastal

central inland, and deep South.

In addition to the description of BlackEnglish in the South,

we also need comprehensive descriptive studiesof nonstandard

Southern white dialects.Although dialectologists have givenus

some indication of the phonology and lexicon ofSouthern white

speech, the grammatical structure ofSouthern white nonstandard

speech is lacking.As was suggested for the study ofBlack English

in the South, several areas shouldbe included, representing

Appalachia, the deep South, and Atlanticcoast regions. Descriptive

studies of this typecan help us resolve the controversyover the

exact relationship between the speechof Southern whites and Negroes

of comparable socio-economic classes.

Descriptions of the correlation of socialand linguistic variables

have also focusedon Northern metropolitan areas.But there are im-

portant reasons why these should beextended to cover severalareas of

the South. Some Southern regional features haveapparently only taken

E .

- 45

on social significance in the North because of their association with

ethnicity and social class in the North.By contrast, there are other

features which have social significance regardless of the geographical

region (a distinction between what I have called "general"and "par,.

tioular" social significance). Careful studies of the social

significance of linguistic variables in the Southcan help us sharpen

our understanding of the interaction of geographical and social factors

in speech.Furthermore, such studiescan lead us to general con-

clusions about the nature of sociolinguisticvariation in the United

States.

Another area of great importance fordescriptive studies is that

of age-grading. The importance of observingage levels in speech

variation was brought out by Hoekett (1950)some time ago but

the actual amount of descriptivestudy has been sparse. Recently

Stewart (1968) and Dillard (1967) havesuggested that an accurate

picture of the nature of Black Englishcannot be studied apart from

a description of age-grading within the Black community.Loban's

(1966) longitudinalstudy of children in California hints atcrucial

age differences, but his taxonomy and linguistic orientationwould

be unacceptable from the viewpointof the linguist.

Studies of age-grading should not beconfused with trite description

of language aoquistion, whichis an area for descriptive studies inits

own right, as we shall shortly see. Age-grade studies should start with

the earliest post-acquisitionalperiod (6-8). The age level when sen-

sitivity to the socialconsequences of speech behavior starts to

approximate adult norms (according to Labov (1965b:91),thisis about age

14 or 15) is of extreme importancefor the linguist. 46 :

The speech of teen-agers is, ofcourse, simply one aspect of their

behavioral response to the adultworld which can giveus invaluable

sociolinguistic information. Such studies, though, cannotbe separated

from peer group norms,so that such studies must concentrateon peer

b4m7.4:93.

Acquisitional studies of nonstandarddialects are also needed for

cross-cultural investigation.But such studies must berelated to

nonstandard norms,a condition which some acquisitional studieshave 16 not observed. For example, the acquisition off and 0 by speakers

of Black English must berelated to the function ofthese units within

the vernacular (e.g. /f/in final position being theBlack English

adult norm).To do otherwisecan only lead to some of the fallacious

conclusions thatve have already dlscussed in PartI of this report.

This is not to say thata cooperative study of nonstandard speech

patterns and certain stagesof acquisition for StandardEnglish

speakers should not beundertaken. In fact, we need such studies

to show us of the ways in which nonstandard dialectsare similar and different from certain stagesin language acquisition.For example, we observe that copula absenceoccurs in Black English and also at a certain stage of languageacquisition for all children,ow we observe that the fi0contrast, one of the lastphonological contrast to be acquired by standard English speakers, ischaracteristic of

Black English in certainpositions.lie need to know in precisely what ways these featuresfunction similarly andin what ways differently.

16. Only by relating it tononstandard no can we have some indication of actual languageretardation bya small minority of lower socio-economicclass children. Current studies which utilize standard English no of acquisitiorerro- neously categorizea majority of these childrenas being linguistically retarded (acase of misconceived retardation). - 47 -

Such studies must serve as the basis for disputing claims that Black

ish indicates a relation to retarded Standard English language

acquisition.

Finally, we need more descriptive dataon the role of sex in

language.Most laymen will readily admit to differences in speech

related to sex, but few, comprehensive studies have dealt with the

topic (perhaps due to our failure to view the familiaras unfamiliar).

The studies by Fisher (1958), Shuy, Wolfram and Riley (1%7),Labov

(1966) and Wolfram (1969)give evidence that this is a quite fruitful

area for descriptive sociolinguistic studies in the United States,

but we need several exhaustive studies showingus the exact ways

in which sex differentiation conditions speech behavioracross

different social groups.

3. Theoretical Issues

Although the explication of theoretical issues isinevitably

related to descriptive studies,we maycautiously isolate several

outstanding issues which current researchon social dialects

raises for the linguist. There are, of course, many issues which

present studies are also raising for sociologists (e.g.the

discreteness of social classes, definition ofsocial roles, etc.)

and/or anthropologists, butin this discussion we shall limitour-

selves to those problems which deal with centralissues in theoretical

linguistics.

Perhaps the outstanding problem for the linguistdealing with

social variation in lanzusge is theway in which observed linguistic variation can be accounted for ina linguistic model of description.

Linguistic models of language descriptionare all based on discrete - 48 -

oppositions visvis gradience or probability.The question s then,

is, do we adopt linguistic models to account for systematic variation

(i.e. variation conditioned by socialor independent linguistic

variables), do we "manipulate" the data in sucha way as to fit into

the existing framework of linguistic descriptions,or do we describe

it apart from any descriptive model of language competence-- a

particular type of performance model?

Iabov has suggested that regular and uniform structuring of variation is an integral part of language competence whereas Decamp

(1969: 1) has insisted that Labov's gradienceis an empirical observation of superficial phenomena whichcan be accounted for by a

"combination of discrete oppositions (cf. Fourier analysisof wave phenomena) followed bycurve smoothing". Despite DeCamp's somewhat cavalier dismissal of Labov's contention, one must recognize the potential that Labov's variable rule has for linguistic descriptions.

Further experimentation with this concept has important implications for the assumption of categoricality in current linguistic models.

With reference to current models, we must also ask if there isone current generative model (e.g. transformational versus stratificational) in which gradience can be incorporated more economically than another.

Such a question may give us some indication of the explanatory adequacy of grammatical descriptions. We must also investigate to what extent descriptive models hold potential in accounting for other types of structured _acial factors conditioning languave choice.Can we, for example, expect and/or demand that a linguistic model incorporate context sensitive rules whose environment is stated in terms of extra'- linguistic factors. Such types of questions that are raised by sociolinguistic investigationscan cause the linguist to

reexamine his assumptions concerning languageas CODE and BEHAVIOR.

Another area in which current sociolinguistic researchmay affect

theoretical models of languagedescriptionconcerns the extent to which a description can encompass more than an idiolect. The traditional approach of the linguist is to describe thelinguistic competence ofa single speaker as representative of a given variety ofthe language.

Certain attempts to accountfor dialect differencesby the incorporation

of "correspondence" type formalization have been tried (e.g.Cochrane's

attempt to formalize Weinreich'sdiasystem (1959), but their focus

on surface realizations make suchattempts descriptively inadequate.

Recently, in the work ofBailey (1968, 1969),a more rigorous

attempt has been made toaccount for different varietiesof a language from a generative view of language.Bailey has proposed that itis possible to give one underlying representation forall dialects of a given language, the difference .between dialectsbeing manifested in the applicability/non-applicabilityof certain rules.

Bailey's "pendialectal" grammar would have rules in their leastgeneral form and their marked order, since themore general forms and theun- marked order could be predicted from the other.The first questionwe must investigate is thefeasibility of suchan approach for social dialects.And, if such an approach is justifiable, what aboutvarieties of English where different underlying structures wouldbe motivated on independent grounds? Does one sacrifice independent motivationfor "overall" descriptive adequacy or are such varieties de factoexcluded as different languages? Ultimately, future descriptive statementsof social dialects which dealwith this mattercan give the linguist information about the natureof dialect differenceswith respect to the surface and underlying forms of a language.

A further area for the theoretical linguist deals with descriptive

differences between several types of language situations.As was

mentioned earlier, there are certain apparent similarities between

standard English- acquisition levels and nonstandard forms..'

It may also be noted that in language pidginization, certain modi-

fications in a language may arise which also show similarities to

levels of acquisition.Furthermore, in language interference of

certain types there is an approximation of some adaptations that

take place in pidginization.Assuming (and only descriptive studies

can tell us if our assumption is correct or not) that there are

similarities between these many types of language "modifications", it does not appear that such similarities would be accidental.What we must ask, then, is if there is something inherent within a language

system which "predisposes certain types of features for modification" in situations such as acquisition, pidginization and interference.

To what extent may we generalize and say that certain aspects of a language are predisposed for modification? (e.g. because of "re- dundancy" or "functional load")? What aspects may be univeral and what ones language-specific? The relation of research to this theoretical problem may give us important clues to universals of social dialects in relation to linguistic structure.

In addition to the broad theoretical issues raised above, there are more specific issues which future sociolinguistic research can help answer. Several of the outstanding issues are as follows: -51 - (1) What is therole of social factors in historical

language change? What implicationsdo they have

for speeding up and retarding changeand how do

such processes operate? (i.e. Furtheranswers to

the problem, Labov (1966) and.Anshen (1969)

attacked).

(2) How does dialect mixturebetween social dialects

contrast and compare with inherentvariability within

a system.Related to this is the question of how

overlapping systemsmay operate in a speech

community or withina single speaker.

(3)How does hypercorrection relateto the linguistic

system? That is, to what extentcan the type of

hypercorrection and the extent of it bepredicted

on the basis of the language and social system.

(4) What can the study ofsocial dialects tellus about

receptive and productive languagecompetence? toes

this apply to all dialectdifferences or only

certain structural categories?Is it reciprocal

between social dialects?

One could go on about the generaland specific implications that future research musthave on current theoreticalissues in linguistics. What ismore important for the linguist, however, is a general approach tosociolinguistics. Sociolinguistic research could simply be understood tomean the description of correlations betwen linguistic and socialfactors, without referenceto any implications that these might havefor theoretical problemsin -52- linguistics. Such studies would, no doubt, have great value for a number of reasons.But for the linguist, sociolinguistic studies have greatest relevance when theyare specifically designed to solve linguistic problems throughan investigation of social factors. - 53 -

21321Laeltty. Anshen, Frank. A sociolinguistic analysisof a sound change. Paper presented at the Linguistic Society ofAmerica, Champaign, Illinois, Jul7:, 1969.

Bailey, Beryl Loftman. Toward a new perspective in Negro English dialectology, American Speech 40,171-177.

Bailey, C.J.N. The integration of linguistic theory: internal reconstruction and comparative methodin descriptive linguistics. yorlatipter_s intics No 2 (1969) 85-122.

. Implicational acalus in diachroniclinguistics and dialectology, WorkingPapers in 1guistics No. 8 (1969) 123-137.

Baratz, Joan C. Teaching reading in the Negro school,in Joan C. Baratz and Roger W. Shuy (Eds.),Teaching Black Children to Read.Washington, D.C., Center forApplied Linguistics.

Bereiter, Carl. Academic instruction and preschoolchildren. In Corbin & Crosby(Eds.),Language programs for the disadvantasd. Champaign, Ill.: NationalCouncil of Teachers of English, 1965. PP. 195-203.

& Ehglemann, Siegfried.Teaching disadvantaged children in the Rreschool. Englewood Cliffs, N.J.: Prentice- 8a11,196-67--

Billiard, Charles, ArnoldLazarus and Raven Mciavid. Dialect features affecting thesocial mobility and economicopportunties of the disadvantaged inFort Wayne, Indiana. "U.S. Office of Education. Contract No. 9-B0114 1969.

Bright, William. Linguistic change insome Indian caste dialect in Charles A. Fergusonand John Gumperz (Eds.), Inguistic diversia in South Asia.Bloomington, Indiana University, 1960.

Introduction: The dimensions ofsociolinguistics, in William Bright (Ed.),Sociolinsaistkes. The Hague, Mouton & Co., 1966.

Cazden, Courtney B. Subcultural differences in childlanguage: An interdisciplinary review.Merrill Palmer Quarterly, 12 (1966), 185-219.

Ghomsky, Noam.Aspects of the theory ofs tax. Cambridge, M.I.T. Press, 1965.

and Mind.New York, Harcourt, Brace& & World, Inc., 1 8:111

Cochrane, G.R. The Australian Englishvowels as a diasystem, Word 15 (1959),69-88.

DeCamp, David. Toward a generative analysisof a post-Creole speech continuum. Paper presented at the Conferenceon Pidginization and Creolization ofLanguage,Mona, Jamaica, 1968. - 54 -

DeCamp, David. Toward a formal theoryof sociolinguistics. Unpublished DS, 1969.

Dillard, Joey L. Age-grading in languageand culture. Unpublished MS, Center for AppliedLinguistics, 1967.

dialect in the innercity, The'Florida . Negro children's FL Reporter 5(Fall1967)97-10.

Ervin-Tripp, Susan 14. An analysisof the interactionof language, (1964), P6-102. topic and listener, AmericanAnthropologist 66:6, Part2

no.3, Language-73ehavior . Sociolinguistics. Working paper Laboratory, University of California,3erkeley, November1967.

theory. Paper. Fasold, Ralph W.A dilemma in-sociolinguistio presented at the LinguisticSociety of America,Champaign, Illinois, July, 1969.

Fischer, John L. Social influences on thechoice of a linguistic variant, Word 14 (1958),47-56.

Dialect Project. Galvan, Mary N. and Troike,Rudolf C. The East Texas Texas, University of Texas,196P.

tapes for changing Golden, Ruth I. EffectivenesP of instructional regional speech. EdD dissertation,Wayne State University.

Gumperz, John J.Dialect differences andsocial stratification 60 (195na), in a North Indian village,American Anthropologist 668-681.

three Hindi village . Phonological differences in dialects, Language 34(195n), 221-224.

the study of Indiancivilization, . Speech variation and American Anthropologist63 (1961), 976-9PR.

interaction in twocommunities, . Linguistic and social American Anthropologist,66:6, Part 2 (1964), 137-153.

continuity, Language Hockett, Charles F. Age-grading and linguistic 26 (1950), 449-457.

Hurst, Charles G. Psychological correlates ofdialectolalia. Washington, D.C.,1965. and Hymes, Dell H.The ethnography ofspeaking, in Thomas Giadwin behavior. William C. Sturtevant(Eds.), Anthropology and human Vashington, D.C., AnthropologicalSociety of Washington,1962 -55-

Hymes, Dell H. Introduction: Toward ethrographies of communication, American Antbropglogist 66:6; Part 2 (1964), 1-34.

Kochman, Thomas. "Rapping" in the black ghetto. Trans-action, February 1969, 26-34.

Kurath, Hans. Handbook of the lksguipticmographE of New England. Providence: American Council of Learned Societies, 1939.

Linguistic ntlas of New England. Providence: American Council of Learned Societies, 1941.

. A word fisgra_silt of the eastern United States. Ann Arbor: Univeri of Michigan Press, 1949.

Labov, William. Phonological indices to social stratification. Paper presented at the American Anthropological Association, San Francisco, 1963a.

. The social motivation of a sound change.Word 12,. (1963b),273-309.

On the mechanism of linguistic change.Georgetown Universi mnamh Series on Language and 18 TT65a,91-114.

. Stages in the acquisition of standard English. In Roger W. Shuy (Ed.) Social dialects and inamat learning. Champaign, Ill.: National Council of Teachers of English, 1965b, PP. 77-104.

. The social stratification of Ehglish in New York Jashington, 11.C.: Center for Applied Linguistics, 39E6a.

. Hypercorrection by the lower middle class as a factor in linguistic evaluation. In William Bright (Ed.), Sociolinguistics. The Hague: Moutons 1966b, PP. 84-113.

. Contraction, deletion, and inherent variation of the English copula. Language .15. (1969).

. The logic of non-standard English, Geortown University Monogranh Series on 1aamel and ansuistics 22lguy:

. Cohen, Paul & Robins, Clarence, & Lewis, John.A study of the non-standard English of Negro and Puerto Rican speakers in New York City. Final report. Cooperative Research Project 3288. Office of Education, 1968.

Legums Stanley E., Williams, Clyde E., & Lee, Maureen T. Social dialects and their implications for buianina reading instruction. Inglewood, Calif. Southwest Regional Laboratory for Educational Research and Davelopreent, 1969. -56-

Loban, Walter D. Problems in oral English. Champaign,Ill., National Council of Teachers of English, 1966.

Loflin, ?Arvin D.A teaching problem in nonstandard Negro English, English Journal 56 (1967a),1312-1314. :... On the structure of the verb dxfardaalect of!...,"imerican Tecro Efiglish. Report l'To.21,Center lor Research in Social Behavior, University of Missouri,1967b. . Negro nonstandard and standardEnglish: Same or different deep structure?Unpublished paper, Center for Research in Social Behavior,University of Missouri, May 1968.

. On the passive in nonstandardNegro Diglish. Journal of MQ1...... sh asa Second Language, 4. (Spring1969)9 19 -23.

McDavid, Raven I., Jr.Postvocalic r in South Carolina.A social analysis. American Smell '23 (1948),194-203.

. A checklist of significantfeatures for discriminating social dialects. In Eldonna L. Evertts (Ed.),Dimensions of dialect.Champaign, National Councilof Teachers of English, 1 376771T. 7-10.

. Variations in standard. AmericanEnglish. Elementarz English 45 (15N7 1968),561-564, 608.

& Austin, William M. (Eds.)Communication barriers for the culturallZ 222Fived. Cooperative Research Project 2107, Office of Education, 1966.

Menyuk, Paula. Sentences childrenuse. Cambridge, BIT Press, 1969.

Nide, Eugene A. Toward a Science of Translatina.Leiden, E. J. Brill, 1964.

Parslow, Robert, Ries, Julie, Bruder,liary,yilson, Timothy and Bessko, Melina. Pittsburgh Dialect Project. Univeesity ofPittsburgh, 1961.

t - .41. Pedereon, Loo A. Tip pronunciation of English inmetropolitan Chicago. publicattot of the 141,:44afMaalectSociety:No'. 44, University; Universityof AlEiblIMi'Pkess, 1965.4-

Schmidt, Paul F. Some criticisms of culturalrelativism, Journal of Philosophy 52 (1955)780-791.

Shuy, Roger W. Wolfram, Walter, A. & Riley,William. Linguistic correlates of social stratificationin Detroit speech.Final report, Cooperative ResearchProject 6-1347, U.S. Officeof Education, 1967.

. Field technigues inan urban la .11.211L. Washington, D.C., Center for Applied Linguistics,19 8.

Skinner, D. F. Verbal, Behavior. New York, Appleton-Century-Crofts, 1957. A

-57-

Slobin, Dan I. A field manual for cross-culturalstudy of the acquisition of communicative competence.Berkeley, University of California.

Stewart, William A.Non-standard speech and the teachingof English. Washington, D.C.,Center forApplied Linguistics, 1964.

. Sociolinguistic factors in the history of American Negro dialects, The Florida FL Reporter5(Spring 1967)11,22,24,26.

. Continuity and change in American Negro dialects, The Florida FL Reporter6(Spring 196P) 3-4,14-16,19.

Weinreich, Uriel. Is a structural dialectologypossible? Word 10 (1954), 3148 -400.

Williamson, Juanita V. A&LoristicaVoll and morphological study ofthe speech of theNemof Memphis. Ann Arbor, PhD dissertation, University of Michigan, 1961.

Wolfram, Walter A.A sociolinguistic description ofDetroit Negro speech. Washington,D.C., Center for Applied Linguistics, 1969.

. Linguisticcorrelates of social differences in the Negro community. Georgetown University MonographSeries on Languages and Linguistics 22 19 9), in press.

. Underlying representations in Black English phonology. Paper presented to the Linguistic Society of America, Champaign, Ill., July, 1969.